2005 Cubs: Run Advancement

[Guest]

There are a lot of different ways to measure the productivity of hitters. One school of thought is to measure their production independent of context. For example, using a linear weights measure the typical single is worth 0.5 runs. Some singles are worth more and others less, but that's the average value of a single. Another way measure a hitter's performance is to look at something called run expectancy and run advancement. Simply put, run expectancy tells you how many runs a team is expected to score in an inning from a given base-out state. For example, MLB teams score about 0.72 runs in an inning after having a runner on second base and one out. Run advancement looks at the difference between when a batter begins a plate appearance and when he ends his plate appearance. Say a batter comes up to the plate with one out and that runner on second. The batter singles to center, scoring the run. The new run expectancy (with a runner on first and one out) is 0.57 runs, so that PA can be said to have a run advancement of 0.85 runs. (RE at the end of the play - RE at the beginning + # runs that scored on the play.) While this method is extremely context-dependent -- meaning it's less valuable for projecting a player's performance -- it is far more accurate at measuring a player's actual contribution to a team's success. because it takes into account the fact that not all singles, doubles, and home runs are created equally.

 

It's no secret that the Cubs underperformed relative to their expected runs scored last year. I thought it may be instructive to look at each player's run advancement to see which players are most responsible for underperforming. All numbers below are relative to an MLB-average player/team:

 

Name                 XRAA     ΔRE    Diff
-----------------------------------------
Derrek Lee           63.7    60.1    -3.7
Aramis Ramirez       23.4    24.5     1.1
Todd Walker           9.9    13.2     3.4
Michael Barrett      10.1    11.2     1.1
Jeromy Burnitz        3.2     5.5     2.3
Jason Dubois         -0.4     2.3     2.7
Ronny Cedeno         -1.2     0.6     1.8
Mike Fontenot         0.2     0.3     0.0
Adam Greenberg        0.2     0.3     0.0
Cliff Bartosh         0.4     0.3    -0.1
Sergio Mitre          0.2    -0.1    -0.3
Jon Leicester        -0.1    -0.2    -0.1
Roberto Novoa        -0.2    -0.2     0.0
Geovany Soto         -0.2    -0.2     0.0
John Koronka         -0.6    -0.4     0.3
Rich Hill            -0.1    -0.4    -0.3
Michael Wuertz       -0.4    -0.5     0.0
Todd Wellemeyer      -0.3    -0.5    -0.3
Ryan Theriot         -1.1    -0.7     0.4
Jody Gerut           -1.7    -1.1     0.6
Nomar Garciaparra     2.3    -1.1    -3.5
Scott McClain        -1.5    -1.5     0.0
Ben Grieve           -0.3    -1.8    -1.5
Kerry Wood           -2.2    -1.9     0.3
Henry Blanco         -3.9    -2.3     1.6
Ryan Dempster        -2.4    -3.9    -1.5
Enrique Wilson       -1.8    -4.0    -2.2
Matt Murton           7.2    -4.1   -11.3
Carlos Zambrano      -0.6    -4.4    -3.7
Jerry Hairston       -6.2    -4.5     1.7
Mark Prior           -3.4    -5.2    -1.8
Matt Lawton          -3.8    -5.4    -1.6
Todd Hollandsworth   -4.9    -5.6    -0.7
Jerome Williams      -5.0    -7.3    -2.4
Glendon Rusch        -5.0    -8.9    -3.9
Jose Macias          -9.0   -17.4    -8.4
Greg Maddux          -9.4   -17.8    -8.4
Neifi Perez         -15.8   -21.5    -5.6
Corey Patterson     -17.7   -28.8   -11.1
-----------------------------------------
Totals               21.5   -33.6   -55.1

Right off the bat you can see how much more accurately run expectancy tracks the 2005 Cubs' run production. According to XR (a typical linear weights measure) the Cubs should have scored about 766 runs last year. (The MLB average was just over 744.) Going by the run expectancy at the beginning and end of each PA they should have scored about 710 runs, far closer to their acutal total of 703.

 

Looking at players individually, it's surprising to see Matt Murton to have the biggest gap between his XRAA and Run Expectancy differential. Going by his component offensive statistics we'd think that Murton gained the Cubs about 7 runs over an average hitter last year, but in fact the team was usually worse off after his at-bats. Before anyone starts protesting or panicking, keep in mind once again that the context-dependent nature of run expectancy means that it's of little predictive value: Just because Murton's hits were ill-timed in 2005 doesn't mean he'll have the same problem for 2006. You can pretty much chalk it up to bad luck.

[Ding Dong Johnson]

By looking at all those numbers, it really seems to me that the problem is I have no idea what the hell I'm looking at.

[Shilzzz]

im probably really stupid, but is that telling me that todd walker was the best player at scoring runs last year?

[sethuel1]

im probably really stupid, but is that telling me that todd walker was the best player at scoring runs last year?

 

Nope, it definitely shows that Derrek Lee was the best at that. What it does show, however, is that Todd Walker "raised his game" more than anyone else when it came to advancing people.

 

BK has just kicked in the "Todd Walker always pops up when there's a man on base" theory in the teeth.

[Guest]

OK, apparently I didn't explain it very well. ;)

 

XRAA stands for eXtrapolated Runs Above Average. It's a measure of how many runs a player produces above and beyond what a normal player would produce in an equal number of plate appearances. It does this by assigning a run value to each offensive event, so a single is worth 0.5 runs, a homer is worth 1.44 runs, a strikeout is worth -0.098 runs, etc.

 

ΔRE means change in Run Expectancy. (Sorry for using a bit of nerdese greek in there, but that triangle symbol is delta, traditionally used to denote a difference.) That's caculated using PBP data. Going back to my first paragraph you see an example of how it's caculated, but here's another one: Say Juan Pierre leads off an inning. With nobody on and no outs, the average team will go on to score 0.55 runs that inning. Thus, the Run Expectancy at the start of Pierre's AB is 0.55 runs. Now consider two different outcomes of that PA; in one he singles to center while in the other he grounds out to second base. If he singles to center field there's a new base-out state of a runner on first with nobody out. From that base-out state the run expectancy for that inning is 0.95 runs. So, that single of Pierre's can be said to have increased the Cubs' Run Expectancy by 0.4 runs. If he had grounded out, on the other hand, the new run expectancy would have been 0.29 runs, meaning Pierre's AB cost the team 0.26 runs. In short, it's basically the difference between how many runs the team is likely to score before a player's at-bat and how many runs it's likely to score after the player's at-bat.

 

What the diff column is telling you is the difference between the player's XRAA and the player's ΔRE. Remember that XRAA is context-independent while ΔRE depends on the base-out state. So XRAA gives the same amount of credit for every home run, regardless of whether or not it was a solo shot or the bases were loaded. ΔRE, on the other hand, will give far more weight to the grand slam. As sethuel1 said, diff can be thought of as a rough "clutch" figure. Players with higher numbers there tended to succeed in high-leverage situations, while players with lower numbers tended to fail when it mattered but succeed when it was less important.

[Lefty]

Ok, I want to see the Cubs record with Patterson and Perez batting 1-2.

[Jon]

Interesting. Thanks for the post, BK.

[Guest]

Ok, I want to see the Cubs record with Patterson and Perez batting 1-2.

The Cubs were 15-15 when both Perez and Patterson were atop the order. Run production was down in those games, though; 4.22 R/G with the duo and 4.36 R/G without. When you look at games with either Perez or Patterson hitting #1 or #2, the Cubs were 54-54 scoring 4.30 R/G with one of the duo while scoring 4.43 R/G when neither of them were at the top of the lineup.

[Bergy92]

interesting, but some people have wayyyy too much time on their hands to be figuring out these things.

[EricG]

Neifi and Corey were worse than all the pitchers.

[sethuel1]

interesting, but some people have wayyyy too much time on their hands to be figuring out these things.

it's called a spreadsheet.

[Guest]

Neifi and Corey were worse than all the pitchers.

More opportunities.

[Ding Dong Johnson]

Neifi and Corey were worse than all the pitchers.

More opportunities.

 

I liked it better the way CubsFanNotBudMan said it. :lol:

[bhogg]

awww, just use batting average and you'll be fine.